In school, I was taught that the speed of light is constant, in the sense that if you shoot a laser off of a train going 200 km/h, it still just goes at a speed of c=299,792,458 m/s, not at c + 200 km/h.
What confuses me about this, is that we’re constantly on a metaphorical train:
The Earth is spinning and going around the sun. The solar system is going around the Milky Way. And the Milky Way is flying through the universe, too.
Let’s call the sum of those speeds v_train.
So, presumably if you shoot a laser into the direction that we’re traveling, it would arrive at the destination as if it was going at 299,792,458 m/s - v_train.
The light is traveling at a fixed speed of c, but its target moves away at a speed of v_train.
This seems like it would have absolutely wild implications.
Do I misunderstand something? Or is v_train so small compared to c that we generally ignore it?
Yes, the speed of light is exactly the same in all directions.
This is a very important and counterintuitive finding, most famously verified by the 1887 Michelson-Morley experiment.
It wasn’t explained until Einstein developed special relativity in 1905.
But it’s not proven to be the same in all directions.
“The “one-way” speed of light, from a source to a detector, cannot be measured independently of a convention as to how to synchronize the clocks at the source and the detector. What can however be experimentally measured is the round-trip speed (or “two-way” speed of light) from the source to a mirror (or other method of reflection) and back again to detector. Albert Einstein chose a synchronization convention (see Einstein synchronization) that made the one-way speed equal to the two-way speed.”
See this for detailed references.
https://en.m.wikipedia.org/wiki/One-way_speed_of_light
The issue being that you’re sending a beam of light to a mirror and seeing the time it takes to take a round trip. It could be faster going to or faster coming back, we don’t know. We just know how fast it went there and back again.
And if you were to send a detector and try to beam light to it, you run into clock synchronization issues - i.e. the time registered on the far away detector is no longer in sync with the one on the point of origin. Want to sync it remotely? Send a signal at the speed of light to tell it the new time, but that depends on the speed of light in that direction, so it may very well remain out if sync. Does the sync happen instantaneously? Claim your Nobel, you just found something faster than light. Plus any sort of issues if you decided to beam the signal back (speed of light, again) from the detector instead of transporting it back and reading it.
Right, yeah, if the whole system is moving in the direction of the light being sent, it would take longer to get there, while the mirrored light would be faster, ultimately zeroing out until its back at the origin.
Sometimes, I feel like we need more blind physicists. Those wouldn’t be horribly confused when they cannot detect things along the way.
That video seems excellent, thanks. Going to have to watch it ten more times, maybe I’ll understand it. 🙃
I do massively prefer it, though, when all the complexity is presented and I have to work it out, rather than a half-accurate, simplified explanation.
the speed of light is constant, in the sense that if you shoot a laser off of a train going 200 km/h, it still just goes at a speed of c=299,792,458 m/s, not at c + 200 km/h.
Yes. It is counterintuitive, but correct.
What confuses me about this, is that we’re constantly on a metaphorical train: The Earth is spinning and going around the sun. The solar system is going around the Milky Way. And the Milky Way is flying through the universe, too.
Do I misunderstand something? Or is v_train so small compared to c that we generally ignore it?
The magnitude of v_train does not affect the speed of light coming in from, or going out in different directions. c simply is constant to all observers, regardless if and how they are moving.
Light emitted or absorbed by a train will always have the speed of c, not c + v_train. Even if that train moves nearly at the speed of light itself.
However, v_train affects wavelength, or color. Light coming in from the front will be more blue, and light coming in from the back will be more red (see ‘red shift’ in the context of distant galaxies). Geometry will warp. But light will always move at the speed of light.
This seems like it would have absolutely wild implications.
Sounds as if the fun begins here. Of what implications are you thinking?
Sounds as if the fun begins here. Of what implications are you thinking?
Well, to massively simplify all of this, let’s assume the Earth is moving in the direction of the North Pole at
c/2.That means, if you’re at the equator and send a signal to a satellite over the North Pole, it’s going to take twice as long as a signal sent across the same distance from the equator outwards.
That feels like it would be quite relevant for GPS.
Ah, yes! Here, we have at least three entities:
- source of light (at equator)
- target satellite (North Pole)
- observer(s)
The crucial question is, do they move relative to each other? If they all move together at c/2, their relative speed to each other is (compared to c: very close to) zero.
So in the context of an application like GPS, you do not want your satellite to smash into Earth at c/2, but have it in a stable orbit. Which means, it’s relative speed is very low. This is why I think the following is a misconception:
if you’re at the equator and send a signal to a satellite over the North Pole, it’s going to take twice as long as a signal sent across the same distance from the equator outwards.
Since both source and target move roughly at the same speed, the signal would take roughly the same amount of time even if the speed of light could be different.
Imagine being in a train and throwing a ball in the direction of movement, and in the opposite direction. As long as you target something in/on the train, it makes no difference. Similarly, the equator light source targeting an orbiting satellite does not care how fast the combined system is moving.
On top of that, the speed of light can not be different for different observers. It is the same for everyone, in each direction, in every motion. This is where relativity deviates from classical mechanics. In classical mechanics, we can make sense of things by imagining interacting billiard balls. Their individual speeds add up when they collide. In relativity, the speed of light is fixed. Nothing adds to it or subtracts from it.
Coming back to GPS: Even at these low relative speeds, there are still relativistic effects big enough to introduce significant errors in GPS positioning. We have to account for these errors: https://en.wikipedia.org/wiki/Error_analysis_for_the_Global_Positioning_System#Relativity
Imagine being in a train and throwing a ball in the direction of movement, and in the opposite direction. As long as you target something in/on the train, it makes no difference. Similarly, the equator light source targeting an orbiting satellite does not care how fast the combined system is moving.
But that example with the ball works, because everything is going at 200 km/h, while the ball is going e.g. at 205 km/h.
My understanding is that light is supposed to work differently. That it does not go at
200 km/h + cwhen fired from a train. Its speed is capped at c.
That means, relative to the train going at 200 km/h, its relative speed should only bec - 200 km/h.that example with the ball works, because everything is going at 200 km/h, while the ball is going e.g. at 205 km/h.
Or 195 km/h. The point is, the ball is moving relative to the train (at 5 km/h). As long as this movement is perfectly constant (no acceleration / deceleration), the overall speed is entirely irrelevant. You can throw a ball the same way at -200 km/h, 0 km/h or a million km/h. Note there is no difference between the system being in motion and being in rest. And you can throw it in any direction, including in movement direction and the opposite. As long as the overall speed remains constant.
Similarly, for the system of equator light source and orbiting satellite, it does not matter how fast this system moves, and in which direction. Or even if it moves at all. So even in classical physics,
lightclassical cannon balls would reach the North Pole and South Pole in the same time.This feature (N/S reachable in the same time) does not change in relativistic physics. It adds another, equally self-sufficient reason: Light has the same speed in all directions, regardless of relative speeds.
light is supposed to work differently. That it does not go at 200 km/h + c when fired from a train. Its speed is capped at c. That means, relative to the train going at 200 km/h, its relative speed should only be c - 200 km/h.
c is not capped at c, but fixed at c. Light always moves at c, for every observer. Regardless of relative speeds between observer and light source, light will always travel at exactly c.
Man, I hate how it sounds like you’re contradicting yourself. As if you’ve somehow gotten extremely confused by my questions or hit your head or something.
Someone else posted this video: https://www.pbs.org/video/pbs-space-time-speed-light-not-about-light/
At least for me personally, that explanation made it click better that we’re not talking about traditional speed, but rather about a general propagation speed limit for causality.I certainly don’t understand the implications yet, but I feel like I just have to think about it and re-read everyone’s responses a few times over.
Thanks a ton for your help!
So is the speed of light constant because light is a particle and a wave? And when the particle moves at a constant speed the change in speed/energie is achived by a modulation of the wave?
I’m going to say that the speed of the train (relative to earth) and the speed of earth (relative to the sun) and the speed of our solar system (relative to the milky way) and the speed of the milky way (relative to space), though it seems immense to you or I, is almost trivial compared to c.
Likely less than 5% of the speed of c, and probably less than 1% of it. It’s unlikely to be anything we can reasonably measure and certainly far less than we would observe without scientific equipment to assist us, it becomes little more than a rounding error.
I agree that it’s probably not significant (I don’t think, we can actually measure it, but if it were anywhere in the realm of c, reality would be very weird).
However, even if it’s at just 0.1% of c, that’s going to have an influence on GPS, where precise timing is crucial.There’s a lot of factors that affect the timing of GPS… It’s built to be robust enough to handle it, and those factors are a lot more notable and common to be able to detect. Also, if there’s an inaccuracy in the speed of light, where it’s biased in one direction or another, it’s something we’ve experienced for our entire existence, and it’s not something we have detected as of yet, at least not in any measurable way by the current technology…
People smarter that you and I combined have worked on this, to great depth, already, and they haven’t found a significant enough issue to even mention it, nevermind publish something about it, and if they had, systems like GPS would have taken that inconsistency into account when designing the system… If they hadn’t, then it’s extremely likely that someone would have noticed when GPS was being made.
With all the intelligent people thinking about problems exactly like this and trying to figure things out for humanity already, I find it’s best to just take c at it’s value, and try not to think about it too much.
I fully understand anyone who does feel like that, but I’m not having it.
From what I’ve gathered in other responses, we simply cannot measure one-way speed of light.You can’t have two clocks, at the start and end of some distance, because you can’t synchronise them. The act of moving them apart could desynchronise them. And the most precise way to synchronise them at a distance is with light, which obviously will not suffice.
So, you need to measure it with one clock, which means returning the light to its origin with mirrors (two-way measurement), therefore negating all directional effects. Well, unless those effects would change enough while the light beam is in flight, but that’s likely near-impossible to simulate.
At some point, people can be as smart as they want, they can’t defy physics to actually measure this stuff. I won’t defy those physics either, I just want to know what’s secured and what’s not…